The comorbidity of alcohol dependence with major psychiatric illness, including affective disorders, is receiving increased attention in the psychiatric literature. However, little systematic research has been done on patients with primary depression and secondary alcoholism. These patients, who typically claim to "self-medicate" their depression with alcohol, are interesting to study because their choice of psychoactive substance may provide a clue to the neurochemistry underlying their symptoms. During this period, we completed a pilot study comparing patients with a history of comorbid primary depression and secondary alcohol dependence to two comparison groups: patients with a history of depression but no history of alcohol abuse or dependence, and patients with a history of alcohol dependence but no affective illness. Patients were matched according to age, sex, and Global Assessment of Functioning score. Eleven patients in each group completed a standardized work-up in which we characterized symptoms, family history, and pattern of drug and alcohol abuse. Results indicate that comorbid patients are significantly more likely than depressed patients to meet DSM-III-R criteria for panic disorder (p=0.04). Compared to depressed patients, comorbid patients have significantly higher hypomania scores (p=0.03) although (with the exception of one comorbid patient) neither the depressed nor the comorbid patients met criteria for bipolar illness. These results indicate that comorbid patients with primary depression and secondary alcoholism may tend to be "trimorbid", with symptoms of depression, anxiety, and alcoholism. In conjunction with the above project, we developed a questionnaire to study patients' use of alcohol, carbohydrates, and caffeine in response to specific depressive symptoms. Test-retest reliability was established. The questionnaire has been administered to normal volunteers and to patients in four separate diagnostic categories: seasonal affective disorder, major depressive disorder, primary depression with secondary alcohol dependence, and alcohol dependence without affective disorder. Patients in the latter two groups did not differ in the reported use and effect of alcohol so that alcoholics without a history of depression were as likely to report drinking in response to depressive symptoms as were those who had had episodes of primary depression. In terms of caffeine and carbohydrate use, the responses of the patient groups did not differ from each other, but all differed significantly from the normal volunteers. Discriminant function analysis distinguished alcoholics from non-alcoholics in the relationship between the alcoholics' reported drinking, anger, and anhedonia.